APPLIED MICROBIOLOGY, March, 1965Copyright © 1965 American Society for Microbiology

Vol. 13, No. 2Printed in U.S.A.

Role of Leuconostoc mesenteroides in Leavening theBatter of Idli, a Fermented Food of India1

S. K. MUKHERJEE, M. N. ALBURY, C. S. PEDERSON, A. G. VAN VEEN, AND K. H. STEINKRAUS

New York State Agricultural Experiment Station, Cornell University, Geneva, New York

Received for publication 2 November 1964

ABSTRACT

_MUKHERJEE, S. K. (Cornell University, Geneva, N.Y.), -'I. N. ALBURY, C.S. PEDERSON, A. G. VAN VEEN, AND K. H. STEINKERAUS. Role of Leuconostoc mesen-

teroides in leavening of the batter of idli, a fermented food of India. Appl. Microbiol.13:227-231. 1965.-The fermentation of the batter of idli, a fermented food of India,was studied. The microorganisms responsible for the characteristic changes in thebatter were isolated and identified. Although there is a sequential change in the bac-terial flora, the predominant microorganism responsible for souring, as well as for gas

production, was found to be Leuconostoc mesenteroides. In the later stages of fermenta-tion, growth of Streptococcus faecalis and, still later, of Pediococcus cerevisiae becomessignificant. The fermentation of idli demonstrates a leavening action caused by theactivity of the heterofermentative lactic acid bacterium, L. mesenteroides. As far as

is known, this is the first record of a leavening action produced exclusively by theactivity of a lactic acid bacterium.

Idli is one of the very widely used fermentedfoods of India, particularly in the South. It isprepared from rice and black gram mungo(Phaseolus mungo), a legume. The ingredients arecarefully washed, soaked in water separately,then ground, mixed, and finally allowed to fer-ment overnight. When the batter has been raisedsufficiently, it is cooked by steaming and servedhot. The product has a very soft and spongytexture and a desirably sour flavor and taste.

Lewis and Johar (1963) reported the species ofLactobacillus delbrueckii, L. lactis, Streptococcuslactis, and a number of yeasts in the fermentingbatter. Radhakrishnamurty et al. (1961) indi-cated the importance of black gram as the majorcontributing component in the fermentation ofthe batter.The purpose of this study was to determine the

nature of the fermentation and the microorgan-isms responsible for the leavening action.

MATERIALS AND METHODS

In the method of preparation of idli there seemsto exist some variation in the proportions of blackgram and rice used. An idli of acceptable qualitywas prepared in the laboratory with black gramand rice in a 1:1 proportion. A higher proportionof rice was avoided, as it induces a starchy flavor.Black gram, obtained from India, with a protein

1 Journal paper no. 1414, New York StateAgricultural Experiment Station, Cornell Uni-versity, Geneva.

content of 24.8%, was used in all the experiments.Parboiled rice of the Uncle Ben trademark wasprocured from local markets. It was found thatbetter results were obtained with parboiled ricethan with milled white rice. It was evident, fromthe preliminary experiments, that the fermenta-tion of rice semolina was much delayed, comparedwith that of the black gram, and that only theground black gram increases in volume when tiletwo ingredients are fermented separately. Theplastic-like or glutinous consistency of the batteris desirable in order to retain the gas evolved.The black gram was washed several times, first

with tap water and finally with distilled water.This removed some of the surface microorganisms.These were found to produce off-flavor in the idliunless they were washed out. The grains weresoaked for 8 hr in distilled water, then ground to afine paste in a Waring Blendor, and mixed withrice semolina (30 mesh) and 1% salt, calculatedfrom the weight of batter. The black gram soakwater was sampled at 0-, 4-, and 8-hr intervals.After the ground black gram and rice batter wasprepared, it was allowed to ferment in a 30 Cincubator. Samples were taken at 4-hr intervals forthe determination of pH and for plate counts ofmicroorganisms. The increments in batter volumewere also noted at those intervals. Total acid andpH were determined in a duplicate batter at 4-hrintervals, so that the batter used for volumedetermination and plating would be altered aslittle as possible.

Microbiological analyses of the black gram soakwater and the batter. The samples taken for micro-biological examination were plated with a Tryp-

227

on April 28, 2020 by guest

http://aem.asm

.org/D

ownloaded from

on A

pril 28, 2020 by guesthttp://aem

.asm.org/

Dow

nloaded from

on April 28, 2020 by guest

http://aem.asm

.org/D

ownloaded from

MUKHERJEE ET AL.

tone-glucose-yeast extract agar, containing A andB salts and Tween 80. (Tryptone-glucose-yeastextract broth was made up of the following: 5g of Tryptone; 15 g of glucose; 2.5 g of yeast extract;1 ml of Tween 80; and 5 ml each of A and B salts,made to 1,000 ml with distilled water. A saltscontained 100 g of K2HP04 and 100 g of KH2PO4per liter, and B salts contained 40 g ofMgSO4-7H20, 2 g of NaCl, 2 g of FeS04-4H20,and 2 g of MnSO4-7H20 per liter. To prepare agar,

15 g of agar were added.) After incubation at30 C for 48 hr, the colonies were counted. Rep-resentative colonies of microorganisms were

isolated and identified. The relative population ofeach species was determined according to themethod of Pederson and Albury (1950).

Gram-positive cocci or short rods, which pro-duced gas and 0.25% or more acid in Tryptone-glucose-yeast extract broth, were tentativelyconsidered as strains of Leuconostoc mesenteroides.Their identification was further confirmed by theirability to utilize the several sugars, and by theircharacteristic growth in sucrose-gelatin stab tubes.L. mesenteroides produces a dextran, characterizedby a mucoid or wooly appearance in the sucrose-gelatin incubated at 21 C.

Gram-positive cocci in short streptococcus

form, which failed to produce gas and producedless than 0.35% acid in the same broth, weretentatively characterized as S. faecalis. Forfurther identification, cultures were transferredinto sugar broths, litmus milk, and 6.5% sodiumchloride broth, and incubated at 32 C. Cultureswere also transferred into Tryptone-glucose-yeastextract broth for incubation at 45 C.

Gram-positive cocci, showing characteristictetrad appearance in acid environment and pro-ducing 0.5% or more acid, but no gas, in the sameTryptone-glucose-yeast extract broth, were ten-tatively considered as strains of Pediococcuscerevisiae. For further identification, cultures weretransferred to sugar broths.

RESULTSThe major changes occurred simultaneously

and reached their peak values within a period of24 hr after starting the soaking of the black grambeans (Table 1). The batter volume reached itsmaximum in 24 hr. The general appearance andthe aroma were then typical of a good idli batter.The highest count was attained after the 20-hrperiod (Table 1, Fig. 1). The greatest change intotal acidity and pH occurred at 28 hr (Fig. 1).

TABLE 1. Bacteriological and physicochemical changes in the idli batter

Estimated no. of each species X 106 Duplicate batter6

Tie Vol of TotalcountMiraoplcseisbatter per g of dry McoephicscespH Total aciditybatter grains X 106 Aerobic pH (g of lacticspecies Leuconostoc Streptococcus Pediococcus acid per g of

mesenteroides faccalis cerevisiac

hr ml

Oa 0.0372 0.015 0.006 0.016 0 6.1 6.2 0.06Ob 0.0117 0.003 0.002 0.007 0 6.2 6.2 0.044b _ 0.183 0 0.174 0.009 0 6.2 6.2 0.128b 120 0 114 6 0 6.25 6.2 0.248c 198 0 188 10 0 6.35 6.3 0.939d 505 326 0 326 0 0 6.00 6.1 0.7712 505 1,260 0 1,170 190 0 5.70 5.75 1.216 645 2,520 0 1,890 630 0 5.20 5.10 2.2320 705 3,000 0 1,500 1 500 0 4.85 4.70 2.5024 745 2,550 0 1,530 1,020 0 4.65 4.55 2.7128 745 4.72 4.50 3.1432 525 2,640 0 1,320 1,190 130 4.70 3.1436 450 2,410 0 846 362 602 4.55 4.35 3.1440 380 2,130 0 852 852 426 4.50 4.30 3.2144 380 1,650 0 990 330 330 4.45 4.30 3.2848 380 1,780 0 980 0 800 4.45 4.25 3.4352 380 2,090 0 420 1,250 420 4.50 4.20 3.4356 380 2,200 0 880 330 990 4.40 4.15 3.5771 380 770 0 460 0 310 4.35 4.10 3.70

460 380 213 0 0 0 213 4.00

a Unwashed black gram wash water.b Washed black gram soak water.e Ground black gram.d Ground black gram, rice semolina, and salt.e Data on a second batch.

228 APPL. MICROBIOL.

on April 28, 2020 by guest

http://aem.asm

.org/D

ownloaded from

ROLE OF L. MESENTEROIDES IN LEAVENING

T75600

500

6

4

PH /% ACID

.............-....

-

* ATTERI VOIUME' -- ACTER AL COUNT

.....--- ACIDITY 2ND BATCH# ---PH

.

0 8 12 16 20 24 28 32 36 40 46 48 52 56

TIME IN HOURS

300oXz

0200 °1

100 xXI

FIG. 1. Changes in batter volume, bacterial count,total acidity, and pH during idli fermentation.

The batter volume, with the air incorporatedduring mixing, measured 505 ml for the batch ofidli containing 75 g of rice, 75 g of black gram,370 ml of water, and 4.7 g of salt. The battervolume was observed to be 380 ml after 40 hr offermentation. The maximal volume of 745 mlwas reached after 24 hr of fermentation. Thebatter did not develop a typical sour aromauntil a period of 16 hr had elapsed, before whichthe aromas of raw bean and rice were apparent.Only after a period of 40 hr was a slight off-aromanoted. There was little change in aroma in stilllater stages of fermentation. Bland flavor and adecrease of sour odor typified the end of fermen-tation. The consistency of the batter was alsosofter, thinner, and less glutinous at this stage.The pH and total acidity changed sharply in

the most active period of fermentation. The bat-ter was fairly well buffered at the initial and finalstages of fermentation (Fig. 2). The acidity wasrelatively high at the stage of consistency suitablefor idli making. At the 20-hr interval, the secondbatch showed a pH value of 4.70 and a total acidof 2.50%, as lactic acid, based on the weight ofdry grains. The pH of the batter of both batcheswas within reasonable agreement after 20 hr.There is an exceedingly high rate of growth of

microbial populations during the soaking andearly fermentation period, rising from a totalcount of 0.0117 X 106 to 3,000 X 106 per g of drygrains, in 20 hr (Table 1). This was followed,however, by a gradual decrease, as shown by apopulation of 1,780 X 106 at 48 hr.A total of 290 bacterial species were isolated

from samples of batter at periodic intervals. Ofthese, 127 produced acid and carbon dixoide, aswell as the typical morphology and mucoidgrowth of L. mesenteroides. Representative strainsfermented glucose, fructose, galactose, mannose,sucrose, maltose, and xylose, with the amount ofacid normally produced by that species. However,they failed to ferment arabinose, and thereforeare considered as arabinose-negative strains of

1 2 3 4ACIDITY

GM LACTIC ACID PER 100 GM OF GRAIN

FIG. 2. Total acidity, pH relationships duringfermentation of idli, second batch.

L. mesenteroides. An additional 93 strains pro-duced acid with no carbon dioxide gas, and repre-sentative strains grew in 6.5% sodium chloridebroth at 45 C; they fermented glucose, fructose,galactose, mannose, arabinose, sucrose, maltose,lactose, raffinose, and mannitol, and were, there-fore, considered strains of S. faecalis. A number ofstrains (56) grew in typical tetrad formation,produced acid but not carbon dioxide, andshowed other characteristics of P. cerevisiae. Theremaining 14 cultures, isolated from the ferment-ing beans, were aerobic, and were not identifiedfurther.The sequence of growth of the organisms is

apparent in Table 1. Some of the colonies (40%)isolated from the original sample of black gramwere found to be aerobes. This situation waschanged rapidly, however, and at 4 and 8 hr ofpresoaking of the beans, 95% of the isolates werefound to be L. mesenteroides. At the 4-hr plating,a few aerobic colonies were still present.The aerobic colonies on the unsoaked black

gram beans are shown in Fig. 3A. The plating at4 hr (Fig. 3B) shows a few aerobic colonies, withthe majority consisting of typical lactic acid-typecolonies. The platings at 8 and 12 hr, as shown inFig. 3 C and D, reveal predominantly the typicalacid-type colonies.The low-acid producers, L. mesenteroides and

S. faecalis, appeared early in the fermentation,followed by the high acid-producing P. cerevisiae.The reason for the greater growth of L. mesen-teroides than that of S. faecalis, a lower acid-producer, in the beginning of fermentation, is notclear. Although L. mesenteroides was predomi-nant during the useful stage of fermentation, S.faecalis was found growing at the start of fermen-tation and comprising 50% of the total count in20 hr. P. cerevisae started growing only after 28hr, a stage when the batter is no longer suitable

VOL. 13, 1965 229

4

I

C

<

on April 28, 2020 by guest

http://aem.asm

.org/D

ownloaded from

MUKHERJEE ET AL.

FIG. 3. Colonies on agar plates during soaking of black gram beans: (a) 0-hr washed beans; (b) 4-hrsoaked beans; (c) 8-hr soaked beans; (d) 12-hr batter.

for the purpose of idli making. Species of yeastsand the species of lactic acid bacteria reported byLewis and Johar (1953) were not isolated at any

stage of the fermentation. The bacterial speciesthey described are homofermentative, and wouldnot produce leavening action.The presence of L. mesenteroides has previously

been established in many fermenting vegetables.Pederson and Albury (1954) showed that, atvarious salt concentrations, the fermentation ofsauerkraut was always initiated by the speciesL. mesenteroides. The fermentation was eventuallycompleted, however, by the species L. plantarumand L. brevis. Pederson and Albury (1956) also

observed the presence of L. mesenteroides in low-salt cucumber brine fermentations. They notedthat L. plantarum was sometimes dominant inone batch, whereas L. brevis was occasionallydominant in another. In all these cases, acid pro-duction was marked.The fermentation of idli demonstrates a leaven-

ing action caused by the activity of the hetero-fermentative lactic acid bacterium, L. mesen-teroides. As far as is known, this is the first recordof a leavening action produced exclusively by theactivity of a lactic acid bacterium. Although thepresence of the homofermentative lactic acidbacterium, S. faecalis, must have significance in

230 APPL. MICROBIOL.

on April 28, 2020 by guest

http://aem.asm

.org/D

ownloaded from

VOL. 13, 1965 ROLE OF L. MESENTER(

acidity development, this organism does not pro-

duce sufficient carbon dioxide to act as a leaven-ing agent.

ACKNOWLEDGMENTSThis investigation was supported by Public

Health Service training grant 5 TI-GM-886 fromthe National Institutes of Health, and by traininggrant 4147 from the Williams-Waterman Founda-tion.

LITERATURE CITEDLEWIs, Y. S., AND D. S. JOHAR. 1953. Micro-organisms in fermenting grain mashes used forfood preparations. Bull. Central Food Technol.Res. Inst. Mysore (India) 2:288.

OIDES IN LEAVENING 231

PEDERSON, C. S., AND M. N. ALBURY. 1950. Effectof temperature upon bacteriological and chem-ical changes in fermenting cucumbers. N.Y.State Agr. Expt. Sta. Geneva, N.Y., Bull. 744.

PEDERSON, C. S., AND M. N. ALBURY. 1954. Theinfluence of salt and temperature on the micro-flora of sauerkraut fermentation. Food Technol.8:1-5.

PEDERSON, C. S., AND M. N. ALBURY. 1956. Varia-tions in the bacterial flora of low salt cucumberbrines. Appl. Microbiol. 4:259-263.

RADHAKRISHNAMURTY, R., H. S. R. DESIKACHAR,M. SRINIVASAN, AND V. SUBRAHMANYAN. 1961.Studies on idli fermentation: Part II. Relativeparticipation of black gram flour and ricesemolina in the fermentation. J. Sci. Ind. Res.(India) 20c:342-345.

on April 28, 2020 by guest

http://aem.asm

.org/D

ownloaded from

634 ERRATA APPL. MICROBIOL.

Role of Leuconostoc mesenteroides in Leavening theBatter of Idli, a Fermented Food of India

S. R. MUKHERJEE, M. N. ALBURY, C. S. PEDERSON, A. G. VAN VEEN, ANDK. H. STEINKRAUS

New York State Agricultural Experiment Station, Cornell University, Geneva, New York

Volume 13, no. 2, p. 227: authors' addresses should appear in footnotes as follows. S. K. Mukherjeeand A. G. van Veen, Graduate School of Nutrition, Cornell University, Ithaca, N.Y.; M. N. Albury,C. S. Pederson, and K. H. Steinkraus, Department of Food Science and Technology, New YorkState Agricultural Experiment Station, Cornell University, Geneva, N.Y. Page 231, "Acknowledg-ments": the following statement should be added to this section. "Both grants are administered by theGraduate School of Nutrition."